Deep neural network to detect COVID-19: one architecture for both CT Scans and Chest X-rays.

Since December 2019, the novel COVID-19's spread rate is exponential, and AI-driven tools are used to prevent further spreading [1]. They can help predict, screen, and diagnose COVID-19 positive cases. Within this scope, imaging with Computed Tomography (CT) scans and Chest X-rays (CXRs) are widely used in mass triage situations. In the literature, AI-driven tools are limited to one data type either CT scan or CXR to detect COVID-19 positive cases. Integrating multiple data types could possibly provide more information in detecting anomaly patterns due to COVID-19. Therefore, in this paper, we engineered a Convolutional Neural Network (CNN) -tailored Deep Neural Network (DNN) that can collectively train/test both CT scans and CXRs. In our experiments, we achieved an overall accuracy of 96.28% (AUC = 0.9808 and false negative rate = 0.0208). Further, major existing DNNs provided coherent results while integrating CT scans and CXRs to detect COVID-19 positive cases.

Shallow Convolutional Neural Network for COVID-19 Outbreak Screening Using Chest X-rays.

Among radiological imaging data, Chest X-rays (CXRs) are of great use in observing COVID-19 manifestations. For mass screening, using CXRs, a computationally efficient AI-driven tool is the must to detect COVID-19-positive cases from non-COVID ones. For this purpose, we proposed a light-weight Convolutional Neural Network (CNN)-tailored shallow architecture that can automatically detect COVID-19-positive cases using CXRs, with no false negatives. The shallow CNN-tailored architecture was designed with fewer parameters as compared to other deep learning models. The shallow CNN-tailored architecture was validated using 321 COVID-19-positive CXRs. In addition to COVID-19-positive cases, another set of non-COVID-19 5856 cases (publicly available, source: Kaggle) was taken into account, consisting of normal, viral, and bacterial pneumonia cases. In our experimental tests, to avoid possible bias, 5-fold cross-validation was followed, and both balanced and imbalanced datasets were used. The proposed model achieved the highest possible accuracy of 99.69%, sensitivity of 1.0, where AUC was 0.9995. Furthermore, the reported false positive rate was only 0.0015 for 5856 COVID-19-negative cases. Our results stated that the proposed CNN could possibly be used for mass screening. Using the exact same set of CXR collection, the current results were better than other deep learning models and major state-of-the-art works.

Automatic Lung Health Screening Using Respiratory Sounds.

Significant changes have been made on audio-based technologies over years in several different fields. Healthcare is no exception. One of such avenues is health screening based on respiratory sounds. In this paper, we developed a tool to detect respiratory sounds that come from respiratory infection carrying patients. Linear Predictive Cepstral Coefficient (LPCC)-based features were used to characterize such audio clips. With Multilayer Perceptron (MLP)-based classifier, in our experiment, we achieved the highest possible accuracy of 99.22% that was tested on a publicly available respiratory sounds dataset (ICBHI17) (Rocha et al. Physiol. Meas. 40(3):035,001 20) of size 6800+ clips. In addition to other popular machine learning classifiers, our results outperformed common works that exist in the literature.

Mechanism of strand exchange from RecA-DNA synaptic and D-loop structures.

The strand-exchange reaction is central to homologous recombination. It is catalysed by the RecA family of ATPases, which form a helical filament with single-stranded DNA (ssDNA) and ATP. This filament binds to a donor double-stranded DNA (dsDNA) to form synaptic filaments, which search for homology and then catalyse the exchange of the complementary strand, forming either a new heteroduplex or-if homology is limited-a D-loop1,2. How synaptic filaments form, search for homology and catalyse strand exchange is poorly understood. Here we report the cryo-electron microscopy analysis of synaptic mini-filaments with both non-complementary and partially complementary dsDNA, and structures of RecA-D-loop complexes containing a 10- or a 12-base-pair heteroduplex. The C-terminal domain of RecA binds to dsDNA and directs it to the RecA L2 loop, which inserts into and opens up the duplex. The opening propagates through RecA sequestering the homologous strand at a secondary DNA-binding site, which frees the complementary strand to sample pairing with the ssDNA. At each RecA step, there is a roughly 20% probability that duplex opening will terminate and the as-yet-unopened dsDNA portion will bind to another C-terminal domain. Homology suppresses this process, through the cooperation of heteroduplex pairing with the binding of ssDNA to the secondary site, to extend dsDNA opening. This mechanism locally limits the length of ssDNA sampled for pairing if homology is not encountered, and could allow for the formation of multiple, widely separated synapses on the donor dsDNA, which would increase the likelihood of encountering homology. These findings provide key mechanistic insights into homologous recombination.

DNA clamp function of the monoubiquitinated Fanconi anaemia ID complex.

The ID complex, involving the proteins FANCI and FANCD2, is required for the repair of DNA interstrand crosslinks (ICL) and related lesions1. These proteins are mutated in Fanconi anaemia, a disease in which patients are predisposed to cancer. The Fanconi anaemia pathway of ICL repair is activated when a replication fork stalls at an ICL2; this triggers monoubiquitination of the ID complex, in which one ubiquitin molecule is conjugated to each of FANCI and FANCD2. Monoubiquitination of ID is essential for ICL repair by excision, translesion synthesis and homologous recombination; however, its function remains unknown1,3. Here we report a cryo-electron microscopy structure of the monoubiquitinated human ID complex bound to DNA, and reveal that it forms a closed ring that encircles the DNA. By comparison with the structure of the non-ubiquitinated ID complex bound to ICL DNA-which we also report here-we show that monoubiquitination triggers a complete rearrangement of the open, trough-like ID structure through the ubiquitin of one protomer binding to the other protomer in a reciprocal fashion. These structures-together with biochemical data-indicate that the monoubiquitinated ID complex loses its preference for ICL and related branched DNA structures, and becomes a sliding DNA clamp that can coordinate the subsequent repair reactions. Our findings also reveal how monoubiquitination in general can induce an alternative protein structure with a new function.

Mechanisms of mTORC1 activation by RHEB and inhibition by PRAS40.

The mechanistic target of rapamycin complex 1 (mTORC1) controls cell growth and metabolism in response to nutrients, energy levels, and growth factors. It contains the atypical kinase mTOR and the RAPTOR subunit that binds to the Tor signalling sequence (TOS) motif of substrates and regulators. mTORC1 is activated by the small GTPase RHEB (Ras homologue enriched in brain) and inhibited by PRAS40. Here we present the 3.0 ångström cryo-electron microscopy structure of mTORC1 and the 3.4 ångström structure of activated RHEB-mTORC1. RHEB binds to mTOR distally from the kinase active site, yet causes a global conformational change that allosterically realigns active-site residues, accelerating catalysis. Cancer-associated hyperactivating mutations map to structural elements that maintain the inactive state, and we provide biochemical evidence that they mimic RHEB relieving auto-inhibition. We also present crystal structures of RAPTOR-TOS motif complexes that define the determinants of TOS recognition, of an mTOR FKBP12-rapamycin-binding (FRB) domain-substrate complex that establishes a second substrate-recruitment mechanism, and of a truncated mTOR-PRAS40 complex that reveals PRAS40 inhibits both substrate-recruitment sites. These findings help explain how mTORC1 selects its substrates, how its kinase activity is controlled, and how it is activated by cancer-associated mutations.

Metyrapone blocks maternal food restriction-induced changes in female rat offspring lung development.

Maternal food restriction (MFR) during pregnancy affects pulmonary surfactant production in the intrauterine growth-restricted (IUGR) offspring through unknown mechanisms. Since pulmonary surfactant production is regulated by maternal and fetal corticosteroid levels, both known to be increased in IUGR pregnancies, we hypothesized that metyrapone (MTP), a glucocorticoid synthesis inhibitor, would block the effects of MFR on surfactant production in the offspring. Three groups of pregnant rat dams were used (1) control dams fed ad libitum; (2) MFR (50% reduction in calories) from days 10 to 22 of gestation; and (3) MFR + MTP in drinking water (0.5 mg/mL), days 11 to 22 of gestation. At 5 months, the MFR offspring weighed significantly more, had reduced alveolar number, increased septal thickness, and decreased surfactant protein and phospholipid synthesis. These MFR-induced effects were normalized by the antiglucocorticoid MTP, suggesting that the stress of MFR causes hypercorticoidism, altering lung structure and function in adulthood.